OpenAI Claims Its AI Disproved the Exponential Time Hypothesis

OpenAI announced that its AI models have found a proof that the Exponential Time Hypothesis (ETH), a 1999 conjecture in computational complexity, is false. The hypothesis posits that the 3‑SAT problem cannot be solved significantly faster than by exhaustive search. OpenAI’s 67‑page paper, dated October 17 and revised October 25, presents a deterministic algorithm that counts all satisfying assignments of 3‑SAT instances in subexponential time, a result verified in the Lean theorem prover. The claim has sparked reactions on crypto‑focused social media, where users mistakenly conflated ETH with the Ethereum cryptocurrency. Independent experts, including Russell Impagliazzo, Ramamohan Paturi, and MIT’s Ryan Williams, are reviewing the proof but caution that independent verification is essential before accepting the result. If confirmed, the refutation would reopen questions about the limits of algorithmic efficiency for a wide range of problems in graph theory, computational biology, databases, and geometry, and could guide future algorithmic research.